TheoremBase

Talagrand's Inequality for a Diagonal Gaussian Measure on Euclidean Space in the Cameron-Martin Cost

A probability measure with finite relative entropy with respect to a diagonal Gaussian measure on Euclidean space is the image of the Gaussian under a Borel map whose transport cost, with each coordinate weighted by the inverse of its variance, is at most twice the relative entropy.

Statement

In the settings of The Real Numbers: Standing Notation and Background and The Intrinsic Calculus on the Wasserstein Space: Standing Notation, with the dimension dd fixed there, the Borel σ\sigma-algebra B(Rd)\mathcal{B}(\mathbb{R}^{d}) of Probability Measures on Euclidean Space and Random Vectors: Standing Notation §spaces, and Borel maps and functions as in Probability Measures on Euclidean Space and Random Vectors: Standing Notation §borel-maps, let c=(c1,…,cd)c=(c_{1},\dots,c_{d}) be a variance vector, let γc\gamma_{c} be the diagonal Gaussian measure with variances cc, and let μ\mu be a probability measure on (Rd,B(Rd))(\mathbb{R}^{d},\mathcal{B}(\mathbb{R}^{d})) with finite relative entropy H(μ ∣ γc)H(\mu\,|\,\gamma_{c}) with respect to γc\gamma_{c}. For a Borel map T:Rd→RdT:\mathbb{R}^{d}\to\mathbb{R}^{d} and k∈{1,…,d}k\in\{1,\dots,d\} write TkT_{k} for its kk-th coordinate function, and T#γcT_{\#}\gamma_{c} for the image measure of claim 1 of that lemma.

(Talagrand) There is a Borel map T:Rd→RdT:\mathbb{R}^{d}\to\mathbb{R}^{d} with T#γc=μT_{\#}\gamma_{c}=\mu such that the nonnegative Borel function x↦∑k=1dck−1(Tk(x)−xk)2x\mapsto\sum_{k=1}^{d}c_{k}^{-1}\bigl(T_{k}(x)-x_{k}\bigr)^{2} is integrable with respect to γc\gamma_{c} and

∫Rd∑k=1d(Tk(x)−xk)2ck γc(dx)≤2 H(μ ∣ γc).\int_{\mathbb{R}^{d}}\sum_{k=1}^{d}\frac{\bigl(T_{k}(x)-x_{k}\bigr)^{2}}{c_{k}}\,\gamma_{c}(dx)\le2\,H(\mu\,|\,\gamma_{c}).

Proofs

Log in to submit a proof.

Loading...

Citations

Loading…

Dependencies

Loading…

Related

0 relations

Curated associations between results. These are editable and subjective — they do not replace the dependency graph, which is derived from the references in the text.

No relations recorded yet.

Comments

Log in to comment.

Loading…